Optimal Sliced Latin Hypercube Designs for Computer Experiments with Continuous and Categorical Factors

• Activity Number: 77
• Type: Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Shan Ba*+ and William Brenneman and William Myers
• Companies: Procter & Gamble and Procter & Gamble and Procter & Gamble
• Keywords: Computer experiment ; Continuous and categorical factors ; Space-filling design ; Maximin distance criterion

Abstract

Sliced Latin hypercube designs (SLHDs) have important applications in designing computer experiments with continuous and categorical factors. However, a randomly generated SLHD can be poor in terms of space-filling, and based on the existing construction method which generates the SLHD column by column using sliced permutation matrices, it is also not easy to search for the optimal SLHDs. In this article, we develop a new construction approach which first generates the small Latin hypercube design in each slice and then arranges them together to form the SLHD. The new approach is very intuitive and can be easily adapted to generate the orthogonal SLHDs and the orthogonal array-based SLHDs. More importantly, it enables us to develop general algorithms which can search for the optimal SLHD efficiently.